In junction-isolated integrated circuits, transient biasing conditions can occur during operation such as to cause the passage of unwanted currents in the substrate of the integrated circuit and within the regions thereof isolated from one another by reverse biased p-n junctions. These currents are due to parasitic transistors becoming conductive. The transient biasing conditions mainly occur upon commutation of inductive loads, such as inductances and motors, or capacitive loads, such as capacitors, batteries and accumulators, effected by means of electronic switches of the integrated circuit.
A typical example of integrated circuit in which parasitic effects of this type occur is a driver circuit for inductive loads, for example, a transistor bridge. Such a circuit is shown in FIG. 1 connected between terminals, indicated +Vcc and the ground symbol, of a dc supply voltage source, which controls a motor M. In this example the transistors, indicated M1-M4 are power transistors of DMOS type, that is, double diffusion MOS field-effect transistors. Each of these transistors has a diode Db1-Db4 intrinsic in its structure, and which acts as a recovery diode. However, a bipolar transistor bridge with reversed diodes between the emitter and collector terminals could be used equally well to describe the parasitic phenomena caused in the integrated circuit by the switching of the inductive load.
As is known, a transistor bridge circuit is controlled in such a way that the transistors in the diagonals of the bridge are alternatively conductive and switched off so that currents in opposite senses are applied successively to the load. The parasitic effects described above occur during switching. Consider, for example, the instant at which the conduction of the transistors M1 and M2 is interrupted before activation of the transistors M3 and M4. In these conditions the energy stored in the inductive load M causes overvoltage in both senses on the output terminals of the bridge to which the load is connected. In particular, the source terminal S of the transistor M2 goes to a voltage greater than that of the supply voltage Vcc and the drain terminal D of the transistor of M1 goes to a lower voltage than ground so that the recovery diodes Db1 and Db2 associated with the transistors M1 and M2 both become conductive.
The effects of the positive overcurrent on the source terminal of M2 are described in relation to FIG. 2. The transistor M2 is formed on a substrate 10 of monocrystalline silicon doped with impurities of p-type, namely in a region 11 doped with n-type impurities delimited by a major or frontal surface of the substrate 10, a buried region 12 strongly doped with n-type impurities and therefore indicated n+, and an isolation region 13 strongly doped with p-type impurities, therefore indicated p+. The buried region 12 and the isolation region 13 form, with the substrate 10 and the region 11 respectively, a pn junction which, in normal operation of the integrated circuit, is reversed biased and electrically isolates the region 11 from the substrate 10. The region 11 provides the drain region of the transistor and has, on a frontal surface, a region 14 strongly doped with n-type impurities and a first metal contact electrode 14' which provides the drain terminal D. A p-type region 15 is formed within the n-type region 11 and provides the body region of the transistor.
A region 9 strongly doped with n-type impurities is formed within the body region 15 and provides the source region of the transistor. A second metal contact electrode 16 is formed on the frontal surface in contact with the source and body regions and constitutes the source terminal S. The source region 15 delimits a channel 17 with the edges of the body region 15. The channel 17 is overlain by a third electrode indicated 18, isolated from the frontal surface by a gate dielectric (not shown) which provides the gate terminal G of the transistor.
In the drawing there is also shown another n-type region, similar to the drain region 11 of the transistor M2, and indicated 11', isolated by a buried region 12' and a junction-isolation region 13, able to contain another DMOS transistor or other components of the integrated circuit. The isolation regions 13 arid 13' of the two n-type regions 11 and 11' delimit a portion 19 of the substrate able to contain other components of the integrated circuit, not shown, for example the control circuits of a DMOS transistor bridge. In this portion of the substrate 10 there is only shown a region 20 strongly doped with p-type impurities which has a metal contact electrode 21 on its surface. This electrode, in the example shown, is intended to connect to a ground terminal, that is, a voltage reference terminal common to all the integrated circuit.
On the other major surface, or back, of the substrate 10 there is also provided a metal contact electrode 8 which is connected to ground. In general, the integrated circuit in the substrate 10 will have several r-type regions, such as the regions 11 and 11' isolated from the substrate by isolation regions such as the regions 13 and 13'.
The body region 15 and the drain region 11 define between them a pn junction which provides the recovery diode Db2 of the transistor M2 in the bridge of FIG. 1. Moreover, the body region 15, the drain region 11 and the substrate 10 define, respectively, the emitter, base and collector regions of a bipolar pnp transistor, represented by its circuit symbol and indicated Qp2 in FIG. 2.
The transient situation described above, that is, where the source terminal of the transistor M2 is at a higher potential than that of the supply Vcc arid the diode Db2 is forward biased, is symbolized by a current generator 22 which injects a current, the recirculation current, into the source terminal S of the transistor M2. In this situation the base-emitter junction of the parasitic transistor Qp2 is also forward-biased so that the transistor Qp2 is conductive and a current is injected into the substrate. Because of the distributed resistance of the substrate, represented by two resistors Rsub1 and Rsub2 in FIG. 2, this current causes a localized rise in potential within the substrate with respect to the ground potential. This can cause disturbances in the operation of the integrated circuit, in particular in the parts in which small signals are processed.
The ground contact formed by means of the region 20 and the electrode 21 provides a known approach for significantly reducing this effect. In practice the effect of the ground contact on the frontal surface is to divide the distributed resistance of the substrate, which is represented by a potential divider formed by two series resistors Rsub1 and Rsub2, the intermediate tap of which is connected to the ground contact 20, 21.
The effects of the negative overvoltage on the drain terminal D of transistor M1 are described in relation to FIG. 3. The structure of the transistor M1 is identical of that of the transistor M2 of FIG. 2 and therefore the corresponding elements are indicated with the same reference numerals. In the drawing various n-type regions, indicated 11" have been shown, similar to the region 11 able to contain other DMOS transistors or different components of the integrated circuit, and a strongly doped p-type region 20 with an ground contact electrode 21 which has the function described above in relation to FIG. 2. The drain region 11 of the transistor M1 provides the emitter region of a parasitic bipolar npn transistor Qp1 the base of which is distributed within the interior of the substrate 10, and, therefore, extends over the whole of the integrated circuit and has various collectors constituted by the various n-type regions 11".
The transient situation described above, that is, wherein the drain terminal D of the transistor M1 is at a lower potential than ground potential and the diode Db1 is forward biased, is symbolized by a current generator 23 which extracts current, the recirculation current, from the drain terminal D of the transistor M1. In this situation the base-emitter junction of the parasitic transistor Qp1 is also forward biased and part of the recirculation current passes through it and is collected in part by the frontal ground contact and by that on the back of the substrate, and, in part (Icp) by the n-type regions 11" which provide the collectors of the transistor Qp1. It can be shown that the ground contacts, especially those on the front, have the effect of increasing the efficiency of the transistor Qp1 and therefore the current Icp collected by the n-type regions 11". To avoid this effect it would be necessary to eliminate the ground contact on the frontal surface, but this would also involve giving up the reduction in the damaging effects of the pnp transistor Qp2 during overvoltages as described above in relation to FIG. 2.